Dados da Submissão

Título

Computer Vision and Deep Learning in Geosciences: Tasks, Methodologies, and Case Studies

Texto do resumo

The use of deep learning tools in computer vision tasks has revolutionized various sectors of society, industry, and academia, simulating human visual capabilities for the interpretation of complex data. In the field of geosciences, these tools have been widely applied in areas such as remote sensing, seismic stratigraphy, digital petrology, among others, facilitating the analysis and interpretation of large volumes of geological and environmental data. This work aims to present the main computer vision tasks that have been utilized in geosciences, including image classification, object detection and segmentation, super-resolution, and artificial image generation. Covering specific methodologies for developing datasets suitable for each task, demonstrating how input data is prepared and structured to train deep learning models. Concrete examples of studies that use these technologies will be presented, highlighting emerging trends and innovations in the field. As these tools advance and become more popular, it is essential to understand their potential and limitations to ensure their effective use and to promote continuous development in the field of geosciences.

Palavras Chave

Deep-learning; Computer Vision; AI in Geosciences; Data Preparation